Camelia, the Perl 6 bug

IRC log for #bioclipse, 2008-11-26

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
06:21 Gpox joined #bioclipse
08:42 CIA-32 bioclipse: goglepox * r7834 /bioclipse2/trunk/plugins/net.bioclipse​.cdk.jchempaint.view/src/net/bioclipse/​cdk/jchempaint/view/SWTRenderer.java: Changed fallback color to magenta
08:42 CIA-32 bioclipse: goglepox * r7835 /bioclipse2/trunk/plugins/net​.bioclipse.cdk.business.test/ (.classpath build.properties): Removed test/ source entry in .classpath and changed source entry in build.properties to point to src.
08:44 CIA-32 bioclipse: goglepox * r7836 /bioclipse2/trunk/plugins/ne​t.bioclipse.cdk.jchempaint/ (78 files in 3 dirs): Moved editor icons to icons/editor and moved the other icons to icons/
08:45 CIA-32 bioclipse: goglepox * r7837 /bioclipse2/trunk/plugins/net.bioclipse.cdk.jchem​paint.view/src/net/bioclipse/cdk/jchempaint/view/ (JChemPaintWidget.java SWTRenderer.java):
08:45 CIA-32 bioclipse: Implemented rendering of AtomSymolElements and come clean up.
08:45 CIA-32 bioclipse: Removed method that rendered text as image and
08:45 CIA-32 bioclipse: flipped it upside down. Initialize settings in
08:45 CIA-32 bioclipse: Render2DModel to draw implicit hydrogens.
08:45 CIA-32 bioclipse: goglepox * r7838 /bioclipse2/trunk/plugins/ne​t.bioclipse.cdk.jchempaint/ (4 files in 3 dirs): Added menu contribution to JChemPaint and implemented decharge handler to use IChemModelRelay
08:45 egonw joined #bioclipse
08:45 CIA-32 bioclipse: goglepox * r7839 /bioclipse2/trunk/plugins/net.bioclipse.cdk.jchem​paint.view/src/net/bioclipse/cdk/jchempaint/view/ (JChemPaintWidget.java SWTRenderer.java): Removed old scaling code from jcp-widget and SWTRenderer
08:46 CIA-32 bioclipse: goglepox * r7840 /bioclipse2/trunk/plugins/net.bioclipse​.cdk.jchempaint/src/net/bioclipse/cdk/j​chempaint/editor/JChemPaintEditor.java:
08:46 CIA-32 bioclipse: Wrapped up propetyChagnge call width Display.syncExec
08:46 CIA-32 bioclipse: solves error related to bug #139
08:46 zarah bug #139 | http://tinyurl.com/54rkap
08:46 CIA-32 bioclipse: goglepox * r7841 /bioclipse2/trunk/plugins/net.bioclipse.​cdk.jchempaint/src/net/bioclipse/cdk/jch​empaint/handlers/DecchargeHandler.java: Avoids NPE when changing charge on an atom from the popupmenu
08:47 egonw moin
08:47 zarah hi egonw
09:19 jonalv joined #bioclipse
09:23 jonalv hi
09:23 zarah hi jonalv
09:23 jonalv egonw: hi
09:26 egonw Gpox: clearly something *did* come up which required my full attention
09:27 egonw should be finished now
09:27 egonw so, will now do that implicit hydrogen thing
09:27 Gpox ok good
09:35 shk3 joined #bioclipse
09:36 egonw hi shk3
09:45 shk3 hi egonw
09:46 shk3 gilliean and me discussed the controller modules yesterday.
09:46 shk3 we came to the conclusion, that the application should register the modules
09:46 shk3 and set the active module (instead of the draw mode).
09:46 shk3 Reasons:
09:46 egonw ah, good of you to come to the same conclusion :)
09:46 shk3 - The app might want to control modules
09:47 egonw even worse... someone might want a vi mode :)
09:47 shk3 - There are some like for enter element, which need a gui component
09:47 shk3 ok, so I will do this
09:48 egonw I think we need an applet version first...
09:48 egonw because that one will stress the design on rendering and controller modules
09:48 egonw much more, than the JCP application and the SWT editor
09:49 shk3 well, the applet is there
09:49 egonw ah, really?
09:49 shk3 just not with a specific gui, but that shouldn't be a big thing
09:50 egonw how large is it?
09:50 shk3 the build system does not yet exist
09:50 shk3 can be run from eclipe only
09:50 egonw ah, ok, but then it does not stress the design yet
09:50 egonw we need the build system first to create these applets:
09:50 jonalv a vi mode for what?
09:51 egonw 1. a 38kB sized applet, with minimal editing functionality (for drawing molecules)
09:51 * shk3 is going for coffee
09:51 egonw 2. a drawing only applet
09:51 egonw 3, large versions...
09:52 egonw we really need to try this 38kB version
09:52 egonw that's our stress case
09:52 egonw Gpox: it seems I need to update cdk.control and cdk.render from jchempaint-primary again, right?
09:53 Gpox yes
09:53 jonalv egonw: net.bioclipse.cdk.jchempaint.view is not compiling for me... :(
09:54 jonalv I sort of need that...
09:54 egonw yeah, same here...
09:54 egonw Gpox: OK, here's the deal...
09:54 egonw I want to upload CDK 1.1.2 into BC2 repos too
09:55 egonw since no API changes happened, it should be ok...
09:55 egonw but might be a bit bumpy
10:18 shk3 hm, don'
10:18 shk3 t see the point with going for the applet now
10:18 shk3 not all functionality is there
10:18 shk3 so testing the size doesn't really work
10:31 egonw yes, it does
10:32 egonw a major reason to have controller modules, is need to pick only the smallest set
10:32 egonw surely, the JCP application could do that too, if you like
10:32 egonw if you plan to do that, you will have to modularize the GUI too
10:32 egonw is that what you intend to do?
10:42 shk3 what do you mean with modularize the gui?
10:43 shk3 "a major reason to have controller modules, is need to pick only the smallest set"
10:43 shk3 yes, and the application needs to decide about it
10:43 egonw change mass number will not be present in the basic applet
10:43 egonw so, neither do GUI elements to do that
10:43 shk3 yes
10:43 shk3 but the gui is defined via gui properties file anyway
10:44 egonw yes, but not the Actions behind that
10:44 egonw those are classes you'll have to remove (modularize)
10:44 shk3 but which classes are in the applet jars will be part of the build process?
10:45 shk3 the code doesn't need to do that?
10:45 egonw don't understand what you mean
10:45 shk3 don't understand what you mean
10:46 shk3 you mean: we need gui elements depending on which application/applet we are in, right?
10:46 egonw yes
10:46 shk3 that's done via the file.
10:46 egonw so, you need clean class depedencies
10:46 shk3 applet size depends on which classes go in the jar
10:46 egonw yes
10:46 shk3 and that's the build process
10:47 shk3 class dependencies?
10:47 egonw more importantly, it's in the design of controller modules
10:47 shk3 actions don't depend on each other (mostly at least)
10:47 egonw the modules define the dependencies
10:47 egonw the build process just follows that
10:47 shk3 no
10:47 shk3 not the applet build process
10:47 egonw not yet, no
10:47 egonw but it will have to
10:48 shk3 why?
10:49 shk3 sorry, let's do that later
10:49 egonw because if a controller depends on the SMARTS engine, the applet will never fit 38kB
10:49 shk3 first we need to get the app running
10:49 shk3 yes, but that' then a controller we shouldn't use
10:49 shk3 in this applet
10:50 shk3 smarts handling isn't basic
10:50 egonw right
10:50 egonw so, what defines basic?
10:50 egonw IMO, that what fits in a 38kB applet
10:50 shk3 more or less, yes
10:51 egonw that's why we need to know what we can scram into a 38kB applet
10:51 shk3 but that has nothing to do with the change in module handling
10:52 egonw why not?
10:52 egonw I think it has all to do what goes into which module
10:54 shk3 but that's not what I am talking about
10:54 shk3 I am talking about how the active module is set in the hub
10:55 egonw yes, setting the active controller should be no problem
10:56 shk3 glad to hear that
10:56 shk3 I am half way through it
10:56 egonw did not mean to critisize that choice...
10:56 egonw I went ahaed...
10:56 egonw splitting up what belongs to a controller module...
10:57 shk3 but the controller module is a class
10:57 shk3 just one class
10:57 shk3 it might use other classes
10:58 shk3 but that will be dealt with while building applet jar
10:58 egonw that would be where the applet story comes in
10:58 shk3 I don't see why we should change that process
10:58 shk3 but let's discuss that later
11:15 shk3 ok, new module handling commited
11:15 gilleain joined #bioclipse
11:25 gilleain hi egonw, is all well?
11:27 shk3 should a EnterElementSwingModule (which opens a swing-based popup) be in jcp or cdk?
11:43 gilleain ack! the META-INF/MANIFEST.MF doesn't exist in n..bc.cdk.jcp.business.
11:46 egonw gilleain: right... that module has been removed
11:46 gilleain ahh.
11:46 egonw please remove it from your eclipse WS
11:47 gilleain and business.test, I presume
11:48 egonw no, that one still exists
11:49 egonw but should no longer depen on jcp.business, but just on jcp
11:49 egonw will rename this to cdk.jcp.test at some later point
11:49 gilleain ok, thanks.
11:49 egonw if you like, file a bug report requesting that name change
11:49 gilleain ok.
12:18 jonalv egonw: are things fixed yet?
12:19 egonw yes
12:19 * jonalv updates
12:20 egonw uploaded CDK 1.1.2 with the Gilleain patch
12:20 egonw 7845
12:22 jonalv egonw: that fix was more like an unfix. Now more stuff is not compiling... :(
12:23 gilleain there should have been a change in the o.o.cdk.sdg plugin
12:24 jonalv libiocml and libiomdl is not compiling for example
12:24 jonalv +what's up with cdk.jchempaint.business?
12:25 egonw jonalv: checking
12:25 egonw gilleain: cdk.layout
12:25 egonw gilleain: that one I ported to the cdk-1.2.x branch
12:25 egonw jonalv: that has been removed
12:25 jonalv egonw: oh
12:26 jonalv that explains some things...
12:27 jonalv what about net.bioclipse.cdk.jchempaint.business.test then?
12:27 egonw I think something did not make it into cdk-eclipse yet
12:27 egonw jonalv: hang on
12:28 gilleain jonalv: I filed a bug report about that
12:29 gilleain @pele gilleain
12:29 zarah [Bug 77] paths in windows slash problems http://tinyurl.com/578aen
12:29 zarah [Bug 78] CR/LF problems on windows       http://tinyurl.com/6eznsa
12:29 zarah Entire list at http://tinyurl.com/674uaz
12:29 gilleain hmm.
12:30 gilleain bug #145
12:31 egonw @pz 145
12:31 zarah http://tinyurl.com/6yb39t
12:31 gilleain right.
12:31 gilleain one day I will get that right.
12:32 egonw yeah, maybe you should write a unit test for that
12:42 egonw http://lauren.vortex.com/archive/000476.html
12:42 zarah egonw's link is also http://tinyurl.com/58alwf
12:44 egonw jonalv: please try 7851
12:45 Gpox 7852
12:45 egonw oh?
12:46 egonw what does that last one do?
12:49 Gpox fixes compile error from renderer api changes
12:50 egonw ok
12:50 egonw good :)
12:53 * jonalv is still waiting for a compiling jchempaint in the bioclipse repo...
12:54 Gpox i was a little fast, 7852 relates to changed in jcp-primary 13290 not in bioclipse repo yet
12:55 jonalv it seems very complicated to use git...
12:56 egonw http://wiki.bioclipse.net/i​ndex.php?title=Updating_CDK
12:56 zarah egonw's link is also http://tinyurl.com/6ag8c7
12:56 egonw Gpox: please load that...
12:56 egonw now editing it...
12:56 egonw so reload every now and then
13:01 egonw Gpox: OK, finished I think
13:08 gilleain Gpox : how do you feel about me changing generation so that a double bond is an ElementGroup(LineElement, LineElement)?
13:08 gilleain (and, similar for triple, and quadruple, of course)
13:09 egonw Gpox, gilleain: how about removing a lot of '2D' from class and method names?
13:09 Gpox i am ok with it
13:09 gilleain Yeah? Ok, it would just mean implementing a RenderingVisitor would be easier.
13:09 gilleain I changed to using AffineTransform, and it turned out to be a bit of a pain otherwise.
13:10 gilleain egonw: sure, that would be great.
13:11 Gpox iirc you moved a lot from x,y to Point2D and back again in your line rendering
13:12 gilleain yes, sorry.
13:13 gilleain I'm still settling on a way to do it.
13:13 gilleain Oh, right, you mean there was a lot of conversion.
13:14 gilleain Yes, it's annoying to go from Point2d->Point2D->int
13:17 egonw I think I am really going to be ill now...
13:17 egonw focusing on anything becomes more and more difficult
13:19 gilleain you're ill?
13:19 egonw on the edge at least
13:19 gilleain :(
13:20 gilleain go and rest, then.
13:20 egonw have meeting in 40 mins which I don't like cancelling :(
13:20 egonw otherwise I would have headed home
13:20 egonw I think
13:21 gilleain but you will go after that meeting, right?
13:22 egonw yes, think so
13:22 egonw but by then it's about time to go home anyway
13:23 gilleain oh well.
13:36 egonw shk3: ummm... addRing(bla, boolean)???
13:36 egonw what about addPhenyl() ??
13:38 gilleain he's gone to a meeting.
13:39 gilleain what has he changed?
13:40 egonw see cdk-devel
13:45 gilleain ahh.
13:50 gilleain stackoverflow error on adding ring. hmmm. wonder if it is related...
13:50 gilleain [:-/]
13:54 egonw git bisect
13:55 gilleain git multiply-entities-beyond-necessity
13:55 gilleain shorthand : git occam
13:55 gilleain :)
13:56 gilleain oh right, now I'm reading the man page for bisect. interesting.
14:36 masak joined #bioclipse
14:49 masak @plan bc2
14:49 zarah masak: bc2/console-egonw-single-line-refactor: 2 h planned
14:49 zarah masak: bc2/console-test-jcterm: 1 h planned
14:49 zarah masak: bc2/managerdocs-finder-html-skin: 1 h 15 m planned
14:49 zarah masak: bc2/managerdocs-improvements: 45 m planned
14:49 masak I pick managerdocs-finder-html-skin
14:50 masak @start managerdocs-finder-html-skin
14:50 zarah masak: Starting task 'bc2/managerdocs-finder-html-skin'. The clock is ticking. :)
14:50 masak aye.
14:53 masak @pause
14:53 zarah masak: Stopping task 'bc2/managerdocs-finder-html-skin' at 3 m out of 1 h 15 m ( 4%).
14:53 masak there's a new error in the syntax parser, due to changes in some file.
14:53 masak @start managerdocs-finder
14:53 zarah masak: Starting task 'bc2/managerdocs-finder'.
14:57 masak ah. it's egonw's new @TestMethods annotation that throws my script off the track.
14:57 masak fixing.
14:58 edrin joined #bioclipse
15:05 masak @stop
15:05 zarah masak: Stopping task 'bc2/managerdocs-finder' after 12 m at 1 h 37 m out of 1 h 30 m (107%).
15:06 masak dang.
15:21 masak @start managerdocs-finder-html-skin
15:21 zarah masak: Starting task 'bc2/managerdocs-finder-html-skin'.
15:39 masak man, I love this kind of work! :)
15:39 masak @plan last
15:39 zarah masak: bc2/managerdocs-finder: 1 h 30 m planned, 1 h 37 m elapsed
15:40 masak @plan current
15:40 zarah masak: Syntax: @plan <task name> <number>h
15:40 masak @plan old bc2
15:40 zarah masak: bc2/managerdocs-finder: 1 h 30 m planned, 1 h 37 m elapsed
15:40 zarah masak: bc2/managerdocs-finder-html-skin: 1 h 15 m planned, 22 m elapsed (active: 19 m)
15:40 zarah masak: bc2/scriptconsole-backspace-bug: 30 m planned, 2 m elapsed
15:40 zarah masak: bc2/scriptconsole-literal-tab-bug: 30 m planned, 15 m elapsed
16:09 masak http://masak.org/carl/managers.html
16:09 zarah masak's link is also http://tinyurl.com/5887bk
16:09 masak @stop
16:09 zarah masak: Stopping task 'bc2/managerdocs-finder-html-skin' after 48 m at 51 m out of 1 h 15 m (68%).
16:09 masak might need a bit of style, too
16:11 masak @start
16:11 zarah masak: Starting task 'bc2/managerdocs-finder-html-skin'.
16:13 masak http://masak.org/carl/managers.html
16:13 zarah masak's link is also http://tinyurl.com/5887bk
16:13 masak ...but I don't see any parameters...
16:14 masak that can't be right.
16:20 gilleain what are you upto?
16:28 masak gilleain: http://masak.org/carl/managers.html
16:28 zarah masak's link is also http://tinyurl.com/5887bk
16:29 gilleain I know..wait, it's changed.
16:30 masak :)
16:30 masak I think I like it now.
16:30 masak @stop
16:30 zarah masak: Stopping task 'bc2/managerdocs-finder-html-skin' after 19 m at 1 h 10 m out of 1 h 15 m (93%).
16:30 gilleain looks good. font size is quite large, maybe that's just me.
16:30 masak yay! :)
16:30 masak gilleain: it is quite large.
16:30 masak I made it smaller and didn't like the result.
16:31 masak but maybe I'll compromise.
16:31 gilleain oh.
16:31 gilleain well, no don't compromise on my account.
16:31 gilleain people can always use cmd-
16:31 masak aye
16:32 masak but I tried 60% before and that was too small
16:32 masak so I went with 100%
16:32 masak tyring 80% now
16:32 masak yeah, that's actually nicer
16:33 gilleain cool.
16:33 masak (generally, I find text on web pages much too small. I'm a frequent user of Cmd++)
16:34 gilleain oh, btw, did you see cdk-city?
16:34 masak gilleain: no. url?
16:34 gilleain gilleain.blogspot.com
16:34 * masak goes to http://gilleain.blogspot.com
16:35 gilleain I can send you the .mse file if you wanted it, but it turns out to be quite easy to generate from within eclipse.
16:35 masak nice one!
16:35 gilleain (and the file is 1Mb zipped :)
16:36 gilleain I actually think that the representation is potentially really useful, especially if a historical analysis was done.
16:36 gilleain Although I think the software would run better if the cdk was written in SmallTalk...
16:38 masak probably.
16:38 masak we should all switch to smalltalk.
16:38 gilleain hmmm. I should like it as a language, but I don't much.
16:39 masak I've never written anything bigger than 10 lines in it, so I don't know.
16:39 masak I think it looks hip.
16:39 masak @stop
16:39 gilleain me either, but I find it a bit cumbersome somehow
16:39 zarah masak: There's nothing to stop.
16:39 masak gilleain: to read, you mean?
16:39 gilleain well, yes, but  to write as well.
16:40 gilleain I think it was the things that ruby has as well | stuff in pipes |
16:40 masak aye
16:40 gilleain that seems like a weak objection...
16:40 masak definitely.
16:40 masak the principle behind those pipes (closures) is very important.
16:41 masak oh, we discussed this at Amazing Thai II! :)
16:41 gilleain yup. clever language features that are rarely used :)
16:41 masak closures FTW.
16:41 gilleain I guess I do like [1..10].each { |i| print i }
16:41 gilleain kind of thing
16:42 masak Smart Girl runs entirely on closures nowadays.
16:43 masak they're part of the architecture.
16:43 masak helps keep the number of parameters down
16:44 gilleain continuations?
16:44 masak gilleain: what about them?
16:44 masak I wish Perl 5 had them.
16:45 masak haven't used them much, but I can definitely see the use in some cases.
16:45 gilleain well, no, not continuations. firstly : what's Smart Girl?
16:45 masak zarah: you're a smart girl, zarah :)
16:45 zarah masak: So smart I'm looking for another job.
16:45 gilleain google tells me there's smartgirl.org "A place for teen girls to share their thoughts" which frankly sounds terrifying.
16:45 zarah gilleain: No hits.
16:46 gilleain oh, you mean zarah.
16:46 masak aye.
16:46 masak I use that name so as not to confuse her :)
16:46 gilleain ahh.
16:47 masak zarah: you're running entirely on closures!
16:47 zarah masak: are you making fun of me?
16:47 masak zarah: yes.
16:47 zarah masak: I see.
16:48 gilleain but, for example : what would one of her closure calls look like? I know about them, and have used them, but I don't see them very often.
16:49 masak gilleain: hold on, and I'll dig up a good example from the svn history.
16:49 gilleain ok.
16:49 egonw going home, bye all
16:52 masak gilleain: http://gist.github.com/29445
16:52 gilleain tack
16:53 masak notice how the sub only takes one parameter now, and the syntax for replying is much more palatable
16:54 masak (the closure in question is the thing returned by the 'reply()' call, from the $dialogue object. it is then invoked to DTRT with the reply)
16:56 masak zarah: you are excused.
16:56 zarah masak: well, yeah. you're excused too.
16:57 gilleain i see.
17:00 masak jonalv, gilleain: http://gist.github.com/29449
17:00 masak the 'sub { ... }' things are closures in Perl.
17:00 masak they close over their current lexical scope, which includes all the variables seen in that piece of code.
17:01 gilleain hmm. yes. I see. I also spot a 'yield', although that's a method, not a keyword, right?
17:01 masak gilleain: yes, it had nothing to do with Ruby's 'yield' or coroutines
17:02 masak it's part of the POE framework, which I use for IRC
17:02 gilleain no, right, it wouldn't make sense in context.
18:33 edrin left #bioclipse
18:34 Gpox joined #bioclipse
19:03 egonw joined #bioclipse
19:03 egonw moin
19:03 zarah hi egonw
19:03 egonw Gpox: ping
19:06 Gpox egonw: pong
19:07 egonw I am not sure how my cold will progress, and not sure I will be at UU tomorrow, or work from home...
19:08 egonw don't think we planned to physically meet tomorrow, but please do not come just to see meet me tomorrow...
19:08 egonw I should make it on Friday though...
19:09 applesauce joined #bioclipse
19:12 Gpox maybe i should come on friday instead then as i would be alone tomorrow
19:12 egonw yes, sounds good
19:12 egonw would be more useful if you were present then, instead of tomorrow
19:15 Gpox i fixed the jcp-manager on my way home the Activator was in the wrong package, and then implicit hydrogens works
19:21 egonw great, thanx

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary